Percy (software)

Percy (software)
Name	Percy
Developer	BrowserStack
Initial release	2016
Latest release	2025
Programming language	JavaScript, Ruby, Python, Go
Operating system	Cross-platform
Genre	Visual testing, continuous integration
License	Proprietary, commercial

Percy (software) is a visual testing and review platform designed to detect visual regressions in web and application user interfaces. It provides automated screenshot capture, visual diffing, and review workflows that integrate with continuous integration services to help development teams maintain visual consistency. Percy is commonly used alongside GitHub, GitLab, Bitbucket, and Jenkins in modern software delivery pipelines.

Overview

Percy offers an automated visual comparison service that captures screenshots of web pages and components across browsers and viewports, then highlights visual changes with pixel-level diffs. It integrates with popular source control platforms such as GitHub, GitLab, and Bitbucket to connect UI changes to pull requests, and with continuous integration providers like Jenkins, CircleCI, and Travis CI for automated runs. The platform is used by engineering and design teams at organizations including Salesforce, Shopify, Atlassian, and Mozilla to prevent unintended UI changes from reaching production. Percy supports frameworks and libraries such as React, Angular, Vue.js, Ruby on Rails, and Django.

History and development

Percy was founded in 2016 by developers focused on solving visual regression testing challenges faced by teams using modern frontend frameworks. Early development targeted integration with Selenium-based tooling and later expanded to support component-driven development workflows originating from projects like Storybook. The company grew its product through releases adding cross-browser screenshots, review workflows, and API endpoints compatible with RESTful API consumers. In 2021 Percy was acquired by BrowserStack, which integrated Percy into its portfolio alongside testing services used by enterprises such as Microsoft and Intel. Subsequent development has emphasized tighter integration with cloud build systems, performance optimizations, and SDKs for languages popularized by web ecosystems, including JavaScript and Ruby.

Features and architecture

Percy’s core features include snapshot capture, visual diffs, baseline management, and review UI. Snapshot capture is implemented via language-specific SDKs that instrument rendering within test suites for frameworks like React and Angular; SDKs communicate with Percy’s API to upload images. Visual diffs are produced by a rendering pipeline that normalizes screenshots, applies anti-aliasing strategies, and computes pixel-level deltas using algorithms influenced by research from institutions such as MIT and Stanford on image comparison. Baseline management supports maintaining golden images across branches in repositories hosted on platforms including GitHub and Bitbucket Server. The review UI provides contextual metadata (commit, author, CI job) and supports approvals or rejections linked to pull requests on GitHub and GitLab.

Architecturally, Percy uses a microservices model running on cloud infrastructure similar to deployments used by Amazon Web Services and Google Cloud Platform, leveraging container orchestration patterns pioneered by projects such as Kubernetes. Storage of snapshots uses object-store paradigms akin to Amazon S3, and the diffing service scales horizontally with job queues inspired by designs from Apache Kafka and RabbitMQ ecosystems.

Integrations and ecosystem

Percy integrates with source control providers GitHub, GitLab, and Bitbucket, CI/CD providers such as CircleCI, Jenkins, and Azure DevOps, and component libraries like Storybook. SDKs and client libraries exist for languages and frameworks including JavaScript, Ruby on Rails, Python (programming language), Go (programming language), and Elixir (programming language). Percy’s ecosystem includes community plugins and third-party tools that connect to design systems maintained at organizations like IBM and Google. Commercial partnerships after the BrowserStack acquisition expanded native connectors for enterprise identity providers such as Okta and Auth0.

Use cases and adoption

Common use cases include preventing visual regressions in feature branches, validating UI changes across responsive breakpoints, and integrating component-level visual testing into design system workflows. Organizations in e-commerce, finance, and media—such as Shopify, Square, and Mailchimp—use Percy to ensure pixel-level fidelity across releases. Product teams pair Percy with accessibility tools like axe-core and performance measurement suites such as Lighthouse to provide comprehensive UI quality gates. Design teams at companies influenced by Material Design and Carbon Design System use Percy to verify component library updates across browsers and device emulators.

Pricing and licensing

Percy operates under a commercial, proprietary licensing model with tiered pricing plans targeting startups, teams, and enterprises. Pricing tiers typically vary by monthly snapshot allotments, concurrent build capacity, and retention policies for stored snapshots. Enterprise offerings include features such as single sign-on via Okta, dedicated support channels, and contractual terms for uptime influenced by service-level agreements similar to those used by Salesforce and Adobe enterprise products. Open-source projects and educational institutions sometimes receive discounted or sponsored access through community programs.

Criticism and limitations

Critics note that pixel-based diffing can produce false positives due to rendering nondeterminism across browsers such as Google Chrome and Mozilla Firefox, or due to font rendering differences from vendors like Adobe Fonts and Google Fonts. Large teams report costs can scale sharply with high snapshot volumes compared to alternatives such as open-source visual regression tools inspired by BackstopJS and testing frameworks used in Cypress (software). Integration with on-premises CI systems like Jenkins or TeamCity can require additional configuration relative to cloud-native CI providers. Finally, some privacy-conscious enterprises raise concerns about uploading screenshots to third-party cloud infrastructure, preferring self-hosted solutions comparable to those offered by GitLab's self-managed runners.

Category:Visual testing tools